Technology and security
Fledium is built for the field: it works where there is no connection, keeps every record with its evidence and shows each user only what they are authorised to see.
Architecture
Four layers. Every entry is stored on the device first and moves to the server when a connection is available; rules, permission checks and AI run on the server.
User interfacesPhone on site, web in the office
- Field app with role-based screens
- Web management panel, on phones too
- Platform console
Offline engineDevice first
- Encrypted on-device entry queue
- Evidence photos kept with the record
- Time-limited, encrypted machine index
Secure cloud serverRules and permissions
- RESTful API, hosted in Europe
- Meter, capacity and location rules
- Permission layer: company, module, action, project
- AI preliminary diagnosis, human-controlled
- Daily reminders and notifications
Database and data securityWho sees what
- A separate data space per company
- Append-only audit trail
- Encrypted evidence store, short-lived links
Works without a connection
In blind spots such as tunnels, mines and remote sites, the field app keeps working. Entries are stored encrypted on the device with their evidence photos. When the connection returns, each entry is sent with its own ID: sending it again never creates a second record, and photos resume where they stopped.
Security
Permissions per project
With role-based access control (RBAC), permissions are granted per company, module, action and project. A record outside a user's scope never appears in lists, searches, the map or reports.
Isolation between companies
Each company's data is kept apart; a user of one company can never reach any record of another. This rule is tested automatically with every change.
Audit trail
With every entry, the staff member's ID, the device details, the GPS location and its accuracy, the timestamp and the evidence photos are stored. Management changes are kept with the old and new value and the reason, in a trail that can only be appended to.
Photos from the camera only
Evidence photos cannot be picked from the gallery; they are taken with the camera. The file's fingerprint is taken on the phone and checked on the server. An entry from a phone that reports a faked location goes to review.
Encrypted evidence store
Entries and photos are kept in an encrypted store closed to the outside; the app opens photos only through short-lived links. Location is captured only at the moment of an entry; there is no continuous tracking.
Photos are not scanned for other purposes
Images are not scanned in the background by AI for health and safety or any other checks; they are stored only as an evidence file visible to authorised users.
Who sees what
Even within one company, everyone sees only their own work. Choose a role and watch what that user would see in the panel and the app change. A record outside the scope is not answered with "forbidden"; it is treated as if it did not exist.
Ready for integration
Today no external hardware or accounting integration is needed. Reports export as CSV; because the server works through a RESTful API, the customer can connect it to these systems if they wish:
- SAP
- Logo
- Vehicle identification systems
- IoT hardware